Lack of knowledge about differential AIDS mortality seriously hampers the study of the economic impact of AIDS in developing countries, at both the macro and micro-economic levels. In this paper, we derive, we think, reasonable assumptions on mortality differentials by age, education, and other micro-economic characteristics by exploiting variables from the Ivorian Demographic and Health Survey. Finally these differ- entials are calibrated on UN demographic projections to obtain disaggragated mortality tables. Our model is based on econometrically estimated equations using commonly available individual characteristics, therefore it can be used to forecast mortality differentials for other surveys as well. A main result of our model is that educated people have a higher risk of dying of AIDS, because they are more likely to have several sexual partners. This effect is however partly offset by a higher probability of condom use relative to less educated people. The identification of the socio-economic characteristics of low and high risk groups seems indispensable to set up adequate AIDS prevention and therapy policies in developing countries.